David Abrahams wrote:
0. Cross-posting to the ryppl-dev list
Lemme register then ;)
1. Yes, in principle it can support those other VCSes. In fact, there's infrastructure in PIP that allows the specifics of the VCS to be abstracted away.
OK
2. My priority is on smoothness of operation. I want all the workflows described at http://ryppl.org/workflows.html to be as painless as possible, and I'm perfectly happy to mandate the use of Git for ryppl if it is going to make the difference between “just works” and “hassle.” If you really want to use Mercurial for your own development you can always leverage hg-git (http://hg-git.github.com), and I'm sure there's something similar for bzr.
It's fine by me. Fact is that I don't *want* to use Mercurial but my surroundings are (university, lab etc) so I'm trying to minimize the problem if I want them to adopt ryppl globally (that's my goal). As for hg-git, i didn't knwo about it, thanks for the link.
3. At the risk of alienating some people… I believe Git is going to win the DVCS wars. It has the momentum. And as truly obnoxious as Linus' remarks about other VCSes have been, I have been working with them too recently, and I am forced to agree with him. So I'm personally not interested in investing much effort in other VCSes.
Let's see. I'm rather agnostic on this side. Whatever do the work properly for me is good.
4. However, if someone comes along who wants to do the work, and can make me confident that s/he'll maintain it, and can make me confident that the system won't have to sacrifice usability, well then I have no objections :^)
It may be this guy we'll end up hiring for our own concern that will have to work wiht ryppl anyway. -- ___________________________________________ Joel Falcou - Assistant Professor PARALL Team - LRI - Universite Paris Sud XI Tel : (+33)1 69 15 66 35